ML20212F971
| ML20212F971 | |
| Person / Time | |
|---|---|
| Site: | Prairie Island  |
| Issue date: | 10/28/1997 |
| From: | Soresen J NORTHERN STATES POWER CO. |
| To: | |
| Shared Package | |
| ML20212F965 | List: |
| References | |
| NUDOCS 9711050186 | |
| Download: ML20212F971 (23) | |
Text
-. . . _ _ . .- . ._ . _ -. ._ _
UNITED STATES NUCLEAR REGULATORY COMMISSION NORTHERN STATES POWER COMPANY PRAIRIE ISLAND NUCLEAR GENERATING PLANT DOCKET Nos. 50 282 50 306 REQUEST FOR AMENDMENT TO OPERATING LICENSES DPR 42 & DPR 60 LICENSE AMENDMENT REQUEST DATED January 29,1997 Amendment of Coolina Water System Emeroency intake Desian Bases Northern States Power Company, a Minnesota corporation, by this letter dated October 28,1997, with Attachment 1 provides supplemental information in support of the subject license amendment request dated January 29,1997. Attachment 1 contains information which clarifies the analyses presented in the supplement dated June 30,1997.
This letter and its attachments contain no restricted or other defense information.
NORTHERN STATES POWER COMPANY By . AM4VV----
Jdsl P. Soretisen Plant Manager Prairie Island Nuclear Generating Plant On this 'l day of Oc 3bN }9 D before me a notary public in and for said County, personally appeared, Joel P. Sorensen, Plant Manager, Prairie island Nuclear Generating Plant, and beir.g first duly sworn acknowledged that he is authorized to execute this document on behalf of Northern States Power Company, that he knows the contents thereof, and that to the best of his knowledge, information, and belief the statements made in it are true and that it is not interposed for delay.
- . ^ ^ ^ :::.::::::,:: :::::::::: a
^
- DALE W.V'NCENT l l notant rvote.w,=ngiota
, DAKOTA CQ U N T Y <
l Y':...^:::7:$$:?-?'jl.',I?',?ll-) )
9711050186 971029 PDR ADOCK 05000282 P PDR
ATTACHMENT 1 SUPPLEMENT 11 to LICENSE AMENDMENT REQUEST DATED January 29,1997 Amendment of Coolina Water System Emeraency Intake Desian Bases Letter to Mr. Don Anderson, NSP, from Mr. Tony A. Kiefer and Mr. William H. Walton, STS, dated October 21,1997 (includes as letter attachments. Figures 1,2, 3, and 4; a " white paper" discussion entitled,
- Mathematical Formulation of SHAKE Transfer Function Versus Deconvolution"; on excerpt from the 1972 SHAKE manual which contains pages 7 and 8 and Figures 3 and 4; and a technical paper by Carl J. Costantino, et al, entitled,
" Seismic Hazard Studies For the High Flux Beam Reactor at Brookhaven National Laboratory".)
3 October 21,1997 h'
hir. Don Anderson Northern States Power Company 1717 Wakonade Drive Welch,hiinnesota 55089 RE:
Response to the NRC Initial Review Questions on STS' Intake Canal Liquefaction Analysis Report Prepared for the Prairie Island Nucicar Generating Plant, Welch, hiinnesota - STS Project No. 28723 A
Dear hir. Anderson:
This letter is in response to questions which were raised by members of the Geosciences Group at the Nuclear Regulatory Commission (NRC) during a telephone conference call held on September 22,1997. The undersigned representatives of SE, our subconsultant Dr. Gonzalo Castro of gel, Dr. A. V. Sctlur of AES, and representatives of your staff par'icipated in the conference call. Our second subconsultant, Dr.1. hl. Idriss was not available for the conference call, but contributed his input to this response.
During the conference call, the NRC described an independent analysis which members of their staff performed using the soil properties, acceleration time history, and soll layering presented in our June 24, 1997 report. The NRC performed this analysis using the CARES computer program which was written by J. Xu, et al. (1990). Reportedly, the NRC enalysis resulted in a computed peak horizontal ground accelerat;on at the free field ground surface of approximately 0.22g and a peak acceleration of 0.15g at elevation 620. These values are higher than our computed values of 0.145g and 0.084g at these levels, respectively. The NRC was concemed about this discrepancy and asked for clarification of the STS analysis. We have paraphrased the main questions asked by the NRC and provided our responses to each question in turn below:
Q.1. Figure 31 in the STS report shows an outcropping DBE motion at elevation 515 with a peak accele:ation of 0.12g. This same hgure shmvs a peak acceleration of 0.084g at elevation 620 within the soil profile. At the ground surface this figure shows a peak acceleration of 0.145g. These values are lower than the independent analysis performed by NRC staff. What is the reason for the apparent deamplification of the acceleration from elevation 515 with a peak value of 0.12g to elevation 620 with a peak value of 0.084g?
A.1. The STS analysis was performed by applying the 0.12g peak DBE time history at an hypothesized outcrop of " competent" material which was defined for the PINGP site as the surface of the weathered sandstone bedrock. An hypothesized outcrop means that the weathered sandstone is assumed to have no overburden cover. The idealized soil profile for the hypothesized oWcrop is shown in attached Figure 1. This procedure was performed in accordance with NUREG-0800, SRP 3.7.1.
In the SilAKE88 computer code, when the motion is entered at the outcropping layer, the program uses a transfer function which computes the base rock motion at elevation 515 with sutt g ng rs r4 1 5 84?2678010 Fan B47?678040
Northern States Power Comp:ny
, S13 Project No. 28723 A October 21,1997 Y 4 "b
Page 2 the free-field overburden soll profile in place above the " competent" layer. The presence of 179 feet of soil above the rock acts to modify the amplitude of the rock outcrop motion. The transfer function from outcrop to within motions computed by SHAKL88 (or other computer codes which include this feature) is always less *han or equal to 1.0 for all frequencies of the motion and is typically lowest at the predominant period (or natural frequency) of the soil profile.
The DBE "within profile" motion at the interface of the weathered sandstone bedrock and the soll profile at elevation 515 has a peak value of 0.08g and is shown in Figure 2. The "within profile" DBE motion was subsequently propagated up through the soil column with the resulting peak acceleration of 0.084g at elevation 620 and 0.145g at elevation 694. Our original Figure 31 was confusing in that it oversimplified the process by not presenting the "within profile" motion at elevation 515.
Since the resulting free-field ground surface motion has a peak acceleration of 0.145g and the response spectrum as shown in Figure 3 exceeds the DBE spectrum, the analysis method is considered conservative.
The mathematical formulation for the transfer function computation, as excerpted from the 1972 SHAKE manual is attached. Additionally, a paper presented by Dr. Constantino, et al, (1991) which discusses (see page 2%) the transformation from outcropping to "within motions" is also attached. This information in addition to NUREG-0800, SRP 3.7.1 forms the basis for the 0.08g peak acceleration at the interface between weathered sandstone and the soll profile.
Q. 2. Does the apparent deamplification shown on Figure 31 result from deconvolution through the soll or through the competent material?
A. 2. It is our understanding that deconvolution is the process of applying a ground acceleration time history (control motion) at or near the top of a soil profile and vertically propagating shear waves downward through lower layers. Convolution is the process of applying the control motion at a competent layer at the bottom of the soil profile and vertically propagating shear waves upward.
Based on this definition, only convolution was performed in the STS andysis. The control motion was placed at the surface of the hypothesized competent material outcrop as previously described. The transfer of the outcrop motion to a within motion at the same elevation is different from deconvolution as presented on the attached sheet titled " Mathematical Formulation of SHAKE Transfer Function Versus Deconvolution". The resulting within profile motion was propagated vertically upward (convolution) from elevadon 515 to elevation 694 as shown on Figure 2.
For informational purposes, STS has now performed a deconvolution analysis in which the DBE control motion was applied at elevation 694 and deconvoluted through a 179 foot thick hypothetical layer (column) of competent material to elevation 515. For this case, the peak Kpnt 28723-A/c123A013. doc l
4 Northern States Power Comp:ny STS Project No. 28723 A m October 21,1997 Page 3 {@
acceleration of the within profile motion at elevation 515 was 0.072g which is less than the 0.08g obtained in the published STS analysis. This is further proof that the transfer from outcropping to within motions is not the same as deconvolution through the competent material.
Q.3. What soll degradation model was used in the STS analysis?
A.3 Modulus degradation curves used within both the SHAKE 88 and QUAD 4hi analyses performed by STS consist of modulus reduction curves presented by Sun et al. In 1988 Appropriate curves based upon confining pressure were assigned to various layers as shown in the SHAKE 88 output presented in our report within Appendix G. The damping modification curves used in the analysis were originally presented in Seed and Idriss 1970. These curves are also shown within Appendix G. The Seed and Idriss,(1970) damping curve was aho modified as a function of overburden pressure. Computed modification factors are shown within the SHAKE 88 output in Appendix G.
Q.4 What is the imped:.nce ratio between the soil and the rock?
A.4 The low-strain shear wave velocity data used within the SHAKE 88 analyses for the site have been summarized on Figure 2. The soil layer immediately above the we lhered sandstone bedrock was assigned a shear wave velocity of 1,950 feet per second (fps). The weathered rock competent material was assigned a shear wave velocity of 2,500 fps. Intact sandstone bedrock below elevation 495 was assigned a shear wave velocity of 5,000 fps. A summary of the initial shear wave velocity and modulus data for each layer within the SH AKE88 analysis is presented within the output shown in Appendix G.
We hope that this information in addition to the attached figures and references helps to clear up the confusion created by our Figure 31.
If you have any other questions or comments please do not hesitate to contact us.
Respectfully, 7 G. z Tony A. Kiefer, P. 2
'nio Project Engineer J N M'w b Q William H. Walton, P.E., S.E.
Principal Engineer cc: Dr. I. M. Idriss, Univ of Ca.
Dr. G. Castro, gel Dr. A. V. Setlur, AES Kpnt 28723-A/c123A013. doc J
Northern States Power Comptny
. STS Pro}ect No. 28723-A 5 October 21,1997 b'.
Page 4 REFERENCES
- 1. Scimabel, P.B., Lysmer, J., and Seed, H.B., " SHAKE A Computer Program for Earthquake Response Analysis of Horizontally Layered Sites", NISEE, December 1972
- 2. Constantino, C., et al., " Seismic Hazard Studies for the High Flux Beam Reactor at Brookhaven National Laboratory", Third DOE Natural Phenomenon Hazards Mitigation Conference,1991.
- 3. Xu, J1, Philippacopoulas, A.J., Miller, C.A., and Constantino, C. J., " CARES (Computer Analysis for Rapid Evaluation of Structures) Version 1.0", July 1990.
h s
Kpr4 28723 A/c123A013. doc J
DBE Outcrop Motien @ Top of Competent Layer 8 g Peak value= ,12 2 1-l 0 .1- M
<t
- 5 l'O Time (secs) 1,
%g l y J
N k k N E
E E-k 7
lj ,
n b
e g )i ANSTEC l [ g u m une CAED e DW
/dco Avnlicblo on ij Z Aporture Card gh 8
IlYPOTilESIZED COMPETENT OUTCROP g p "3 3 (OVERBURDEN REMOVED) b b g u. a: 2 n V// V# V#
20 n. Weathered Sandstone Bedrock V, = 2500 fps g 8O O f
" S m PQ g @ U Sandstone Bedrock 9 Ge 5E V, = 5000 fps 8h g 5 E Gil @
ON Uf e
1 s'.*JMt",i;.'.%
STS PRO E T STS PROJECT FILE AS SHOWN
( 1
Surface Motion @ EL.694
.15 Peak value= .1451 4
} ,3 >
8 oo-o
^}v p"' f(
h' ,6 9 Ny ^=
tfrft ?- --
- .05- N o tI j j E'674
~
~
l O l sL 671
.15-Time (secs)
,33, Motion Within Profilo @ EL.620 st 62o S' .1- peak value= .084
@ .05- $ ,
}
3 o & Mli . d ;) W l d. 4 .
A ~
.05-
'I' h Free Fi<
.15- S!!AKE,
'Colum 5 10 Time (secs)
DBE Motion Within Profilo @ EL.515 rt QL,,
.ss. (SHAKE 88 Transfer Function Calculation
- I' g Peok valuea .0802 c 05-to 0- - ^ v^ yA y - - h --
c l - -
y yy . -
' t E
g .i-
.15-
$ l'0 Time (segs)
Note: Vertical Scale i
- s .,
8 R R R ANSTEC 5 5 5 @
APERTUflE g x g
j CARD o 5 j
Also Avollable on i Anoduro Card I k "
\g\ I
); I 2 PLANT STRUCTURES E L __ $ V, = 650 fps 2 g p]
y
,,,, g ,,,, 9 .
4 4 -
6 j V, = 820 fps 6 M
d (l $
z$
h; "y, thoo fps EI635 h"o l
h f -
D$
/ j) fl y8@!
e" s i ih,;
'i k(8d J Q z
7 f a V, - 1000-1950 fps 55g
-> ' i e B!
O 95 n
A Direction of Shear Wave Propogation (Convolution) ww w- w
,,i 8 Weathered Sandstone Bedrock
$j V, = 2500 fps
?
9 Sandstone Bedrock V, = 5000 sps No$su$0n'h"gYdes q STS PROJECT NO.
28723 A I /(/0 0 -
/ STS PROJECT FILE SCALE O 200 AS SHOWN 5 -8x llorizontal Scale SCALE IN FEET FIGURE NO.
[
-___ _ b.
7 7 7 8- 9- 8-7 7 7 1
1- l s%r 0 0 M 1 1 l
u s.
f n
9a A P K hhpo T W A s D U T nm 0 Y
" Y B S 1
Y B D E B D E L 4 E R 9 5 M K W L 6 C R E 1 A E P C 5 R H P X L ) L D C A E l I i
- E 9
@dB i S @
0 r il ob l e
tu op f
^.
le(c Ms b r
ioM P
r nk 8 e ih ip 0 T
f a-r it o
t N
u M
2 Ws a A \
\
S- ge)c n( L
- nh o P ipI5 it o 7 A G pl b
Ml 0 R N O
. T e- Up I
E C G T
[ dT t
O (3 B
D E N AR A I
P P E T S
E B 6 E M A N EO S D l S
N D G E b -
\
0
)
s O T R P N A N N d: I
> r S E E M p E C L ,
" R R C H 30 s(e D E L P U 5
N C LE d E D o
I F RNW i
r E O A E F L e
P R SI 4 F 0
EI R
h I
A R
" P Z 3 0
- D b
- 2 0
- a. -
u.;, -
t s ;
n .
a A aus % :.
i 1
0 N n o =.
c_
s T
sC
$ 5 0 5 0 5 0 I[ 0 0
5 4 4 3 5
2 5 0 5 0 $ 5 3 0 kE**
1 1 0 0 0 0 0 0 0 0 0 0 0
- 22.3i30k e g
m b l ll !C
CASE A CASE B
- EL 694.0 - * * " - EL 694.0 Ut = 2E t e# Ut = 2Et e #.
A n n 1
h3 Y Et Et
,r Et Et B
Um = (Em + mF )e' a w m hm l5
'f n b l
p b E
Fm Em "
o g -
NN
- EL 515.0 .
- EL 515.0 .
UN=(EN + IN)e VN= 2EN e' Uh = (EN + I~N )* N = 2E~N e' a n n n a n
" l
~ N hN hN FN EN COMPETENT ayEa ,,
EN EN IN h h h 3r 1r HYPortO1 CAL e e e A) SYSTDA ANALYZED B) COMPETENT LAYER DECONVOLUIiCN THROUGH OUTCROPPING COMPETENT MATERIAL 5 6 S E ""
R R ;w SHAKE TRANSFER FUNCTION KKB 10-13-97
~
z
- 5 i% g $;?
5[~uy
% VS DECONVOLUTION
-e TAK m
10-13-97 A PRA!RIE ISLAND NUCLEAR GENERATING PLANT "* wgw A >P srs con.unont ua.
WELCH, MINNESOTA -
3o_33_g7 Consumag Engineers
Mathematical Formulation of SII AKE Transfer Function Versus Deconvolution a
CASE A: Analyzed Case Control Motion Applied to Rock Outcrop at the Top or the
] Competent Layer '
1 ,
Following the theoretical development for the solution of the wave equation, as given in ,
Section 2 of the SHAKE Manual (Ref.1), the ratio of the motions, as a function of frequency, 4
between the outcrop layer where the control motion is applied and any other layer within the soil profile, including the interface of the competent layer and the soll profile is determined by
- the incident and reflected wave components at those layers. Of particular interest is the ratio
! between the within profile motic.n at the top of the weathered rock (i.e. at El. 515) and the motion at an hypothesized outcrop of the weathered rock. This ratio is given by (Ref.1):
)3 A,y (m) = ny = y (m) + fy (m) c !
j n'y 2cy(m) 1 where, un and u's are the motions at the top of the within profile layer at El. 515 and at the hypothesized outcrop (where the control motion is applied) respectively. This is shown schematically in Figure 4. The terms en (m) and fu (m) are the incident and reflected wave 4
3 - components of the motion at layer N where the outcrop is hypothesized to occur, t
4 Physically, the above equation shows that the within-profile motion (numerator) is composed of two components: the incident wave component en which is not a function of the properties of the layers above it, and the reflected wave component fu which is dependent on the
- overburden soll profile. The denominator is represented by 2 cu because at the outcrop the incident motion is completely reflected since it is a free surface. If the control motion (i.e. the DBE expressed in the frequency domain) is represented by, DBE(m), then the within-profile
, motion at El. 515, ACC'm , will be given by :
l ACC'm(s) = A'y (m)
- DBE(m) i Therefore, the motion within the profile and at the hypothesized outcropping for the same layer l will be different, with the within-profile motion having smaller amplitudes at all frequencies a
because of the overburden, as can be seen in Figure. 3. Note that the within-profile motion at El. 515 contains, or is affected by, the incidence and reflections of waves through the soil profile above 515 Similarly, the motion at the profile surface is given by :
1 ACC^m = A', (m)
- DBE(m),where 1
, A'i (m) cy = -(m)
, A's (m)is the transfer function from the outcrop motion
- to the surface of the profile motion Kproj. 28723-A/c123A012. doc
..-_-.-_.d,--- , . . . - _ _ _ - - - , , - - . _ . . -
I CASEB: Alternative Case - Control Motion Applied at the Smface of a 11ypothetical Column of Competent Material and Deconvoluted Deconvolution of the control motion applied at the surface (El. 694) through an hypothetical column of the competent layer is shown on Figure 4, Case B. Note that the same notation is used as in Case A for consistency.
The ratio of the motion at an hypothesized the outcrop of the weathered sandstone at El. 515, to the surface (control) motion is given by:
jn., (g) , It"*# , c"x (m) 1 1 The acceleration at the outcrop at El. 515 is given by:
ACC*mewar = A"*x (m)
- DBE(m)
This motion is different than the control motion which was used in Case A at the same location and is dependent on the hypothetical single column competent material properties.
The ratio of the within profile motion at El. 515 to the control motion applied at the surface (El.
694) of the hypothetical column is given by:
- A" . (m) = """ = #"" (*) + I"" (*)
1 1 Note that the surface motion at El. 694, is represented as 1 since the functions, e and f, are obtained with respect to the surface (Equation 20 and 21 of Ref.1). The motion within thc hypothetical single column at El. 515 is given by:
ACC*m = A"x (m)? DBE(m)
Again, this motion will differ from the Case A within-profile motion because of the application point of the control motion and the difference between the actual soil profile and the single competent material column.
Kproj. 28723-A/c123A012. doc
4 #
k EARTHOU. . ENGINEERING RESEARCH CEN'iER BHAKE A COMPUTER PROGRAM FOR EARTHQUAKE RESPONBE ANALYSIB 0F HORIZONTALLY LAYERED SITES by Per B. Schnabel John Lysmer H. Bolton Seed A Computer program distributed by NIBEF / Computer Applications Report No. EERC 72 - 12 December 1972 e
College of Engineering University of California Berkeley, California 1
ti(x,t) = 8'y = - w'(sei M *) + Fe-i(kx-we)) (23) at and strains by:
y= = ik(Ee I ") - Fe ~") ) (24) 2.2 Ratio between rock outcrop motions and base rock notions.
If the amplitudes of the incident and reflected wave components, Eg and Fy , in the elastic halfspace, Fig. 3a, are known, the motions in the halfspace with the soil system removed, Fig. 3c, are easily computed. The
- shear stresses are sero at any free surface; thus F = N E , gand the incident wave is completely reflected with a resulting amplitude 2E atN the free surface of the halfspace. The amplitude of the incident wave in the
{
4
! halfspace.is independent of the properties of the system above it since 4
l the reflected wava is completely absorbed in the halfspace and does not i
contribute to the incident wave. The incident wave component, E y, is l
j therefore equal in all systems shown in Fig. 3.
l The ratio between the base motion,N u , and the a tion, uj, at i
! the free surface may be computed from the transfer functiont u, ey (w) + f,(w)
(25)
N(W) " { " 2g(w)
The transfer function between the motion at the surface of the deposit, u y, j and the motion at the free surface of the halfspace ist
- g,1(w) - y , (a) .
l If the halfspace is the rock formation underlying a soil deposit, Eq. 25 shows the ratio between the motion in the base rock and in the out-cropping rock. The ratio between tho'ang11tudes of the base rock action I
a
- ,n - - _ , - - . ,r,-,
I .'.
/ is' u, = 2E,e #
a n I h,
[
IAt s i IAf
. u, = (E,+ F,)e u, r 2E, e U
a n n
< r F, E, E' E' m h, m h, o n , iAt u, = MN+ FN )* *
"N* Ne
<r __
o ; __ ;
N m F E E E E N N N N N , ,N ,r u
System analyzed Layer m outcropping Holfspace outcropping Fig. 3 ONE- DIMENSIONAL SYSTEM WITH OUTCROPPING LAYERS i
I
. [' g I and the outcropping rock notic , is always less than 1, with minimum values at the resonance frequencies of the deposit. Transfer functions for the deposit used in the example,(Sect. 6), are shown in Fig. 4. The amplitude of the base rock motion is only 65% of the amplitude of the rock outcrop motion at the fundamental frequency of the deposit. This dif ference is a function of the impedunca ratio between the deposit and the rock and of the damping in the deposit.
The differences in the computed responses resulting from the use of a rigid base, relative to the use of an elastic base, depend also on which frequencies are dominant in the rock motion. Rock notions with frequency dominance near the resonant frequencies of the deposit will be considerably more affected than motions with frequency dominance between the resonance frequencies, see Fig. 4.
The effect of the elasticity of the base rock is, therefore, not only a function of the impedance ratio between deposit and rock and of the damping in the deposit, but also of the frequency distribution of the energy in the rock motion relative to the resonance frequencies of the deposit.
An approximation for the free surface motion for one of the layere in the system, Fig. 3b, may be obtained in the same way as for the halfspace, provided the incident wave component in the outcropping layer and in the layer within the system are equal-i.e. E, = E' . This is approximately the case when the properties of layer a and all layers below are equal in the two systems and when the impedance, p, V,, is of the sans order of magni-tude _ as for the halfspace. This is the case for exagle, in sedimentary rock layers overlying a crystalline rock base. For a more accuate colution, the motion in outcropping layers must be computed in a separate system from the motion in the halfspace.
< i l1l ,
~
.I 8 I 7 A _
. I 6 A
e s
a S
b
. N k
O I
c T o .. I 5 C r %. .
N U
N, o c
._ F
, t I
A
'A e p s o ' / R e
e r c c E c /
c t f
a r
f a
r u
u o -
I 4 (F c
S N
u s k n s c e A l
i o - u R l
i o R -
q T o s e s .
o - r 4
t I
3 F o .
t b -
g k
o r
i c c .
F o
r t
u t 1 i o
e k .
I 2 s c -
a o B R I
f p1 t l I II \
I 1II II I IIl I I
l
,r I I f
- - O OI 5 1 O
coI*{1E<
f I < l
. SEISMIC HAZARD STUDIES FOR THE HIGH FLUX BEAM REACTOR AT BROOKHAVEN NATIONAL LABORATORY Carl J. Costantino and Ernest Heymsfield Civil Engincedng Department, City Univenity of New York New York, NY 10031 and Young J. Park and Charles H. Hofmayer ,
Structural AnalysisDivision BrookhavenNationalLaboratory Upton, NY 11973 .
ABSTRACT This paper presents the results of a calculadon to determine the site specific seismic hazard appropdate for the deep soil site at Brookhaven National Laboratory (BNL) which is to be used in the riik assessment studies being conducted for the High Flux Beam Reactor (FSBR) ' Die calculations use.
as input the seismic hazard defined for the bedrock outcrop by a study conducted at Lawrence Livermore National Laboratory (LLNL), Variability in site soil properties were included in the calculadons to obtain the seismic hazard at the ground surface and compare these results with those using the I generic amplificadon factors frotn the LLNL study.
INTRODUCTION column by means of these generic amplification factors may not be appropriate for any particular A numerical procedure has been developed to sitein question, determine site specific seismic hazards appropdate for the deep soil site existing at BNL from To accomplish this objective, a Monte Carlo compuable hazards defined at the rock outcrop. procedure has been developed which includes the-The input rock site hazed definition was obtained effects of vadability in the 7arameters descdbing from the site study performed by LLNL (1). The the properties of the soil overburden which rock outcrop hazard is specified in terms of the influence site response. This procedure has been annual probability of exceedance of the peak applied to the deep soil site existing at HFBR. A ground acceleranon (PGA) and the assoctated convolution method of ac.alysis is used, assaming uniform hazard spectra (URS) corresponding to upward propagating sher.r waves, to convert rock each retum period of interest. Tins data is motions appropnate for the rock outcrop at the site converted in the LLNL study to the corresponding to surface soil responses and corresponding hazard site hazard appropriate to the top of the ground definitions. Variability effects from input rock surface by using generic frequency dependent soil motion, soil shear modull, effective hysteretic amplification fgetors 'Ibese facton account for the damping ratio and strain d ency are included effects of the overburden soils on the seismic in the procedure to de e the surface hazard response and are typically obtained from studies of predicuons.
upward column.However, propagating shear questions of waves throughofthe SITE the magnitudes soil DESCRIPTION these factors have been raised, particularly for deep soil sites such as that at BNL. In several The site description was obtained 'from applications considered to date, it has been found previous studies conducted at BNL as well as from that the conversion of the input rock hazard to the studies of the nearby Shoreham Nuclear Power .
corresponding definition at the top of the soil Plant and is summarized in [2]. A cross-section Third DOE Natural Phenomena Hazards Mitigation Conference - 1991
. 205
through the BNL site is shown in Fig.1. The becomes extremely important. For this depth of the soil overburden at the site is investigaden, seismic motions were calculated at ..
approximately 1550 feet and consists of relatively the ground surface using additional postulates of dense gravels and sands interspened with stiff the nonlinear properties of the foundation soils to ;
clays and sandy clays. Blow count data obtained obtain the sinvity of the predictions to these from standard penetration tests were obtained from assurnpdent several borings taken in the area through the upper 100 feet of sedirnents, with soil des::riptions for the To obtain estimates of soll stiffness required deeper sediments obtained fiom well logs for th: hazard calculations, the number of blows appropriate to the area. No other strength or required to drive the Studard Penetration Test stiffness information was available for these soils. (SPT) sampler (taken at five foot intervals to The vadability in blow count data for the near depths of about 90 feet) was used and converted to surface sous, shown in Fig. 2, was significant and effective low strain (initial) shear modulus. Since can be considered typical for the site. This this SPT data is used directly to estimate the low variability was included in the she specific strain shear modulus, the variability in this data calculations to try to capture the effect of this must be suitably accounted for in the hazard uncertainty in the convoluuon studies, calculations. The SPT data was first modified for the effects of depth by converting to equivalent SrrE RESPONSE CALCULATIONS blows at a standard depth ((7)). This corrected data was then used to obtain bounding estimates of ne site specific response calculations made blow counts for all soils in the column.
use of the standard assumption of horizontal shear waves traveling upward through the long soil The initial soil shear stiffness at any depth in column from the basement bedrock up to the the soil column is obtained from standard ground surface, with nonlinear soil properties relationships for the various soll types. For sandy beine included in each sp ' e culati b the soils, for exarnple, the initial shear modulus is i
tivuntyth_cdsJ each calcu ation, t. e obtained directly from esdmates of relative density input rce.k motions wue specified as outvop and confining pressure where the pararneters are motions applied at the top of :>cdrock. Cornpatible directly related to the relative density of the sands motions within the soil column were then and the confining pressure at depth. *Ihe reladve calculated which suitably' account for reflection and density can in tum be estimated from the SP'r blow refraction effects at the bedrock / soil interface as counts for the soll fic:a a variety of rehtions.
well as at all layer interfaces within the soil However, as may be expected, the wide range of column. The CARES Computer Code (3) was varbility in estimated soil stiffness may be used to perform these calculadons, obtained, depending upon the amcular relationship utilized. This varia was Initially, e so co umn culations were incorporated into the calculation by usin bounds made using the standard Seed-Idriss stmin on these recommendations, with a further randorn dependent soll properties typically used in site generation to select relative density from the SFT evaluations (4). These effects are represented by blow count data. For any sandy soll layer in the the degradation in shear tuodulus and increase in column, random nurnber generato;s were used to bysteretic damping ratio which accompanies estimate first the corrected SPT blow count increased shear strain levels as shown in Fig. 3. associated with that layer between the lower and However, recent studies ((5), (6]) have indicated upper bound values found from the site borings, that the degradation in soll properties postulated in and then the relative density for that parthular blow the original Seed Idriss formulation may in fact be count, he initial soil stiffness could then be too1 e so as to preclude the ability of significant computed directly from the relations mentioned high uency energy from being transmitted above. An addidonal random number generator upward ough the soil column. As may be noted was included to account for additional s%ter in the in Fig. 3, the roodified models indicate available data used to define these parameten, significantly less degradation with strain as cornpared to the original Seed Idriss model. At any For the clay soils at the site, the tehtive deep soil site, such as at BNL, the form of the density wr.s not directly used in the calculation.
Rather, the initial soil stiffness was related to the degadation properties usJ the analyses s
Third DOE Natural Phenomena Hazards Mitigation Conference - 1991 206
o .
inidal void ratio of the soil as well as the column cases were run to obtain surface ground overburden pressure at depth and the motions for each of the degradation models ov=consolidanon ntio of the soil. A nndom considered (100 runs for the Seed Idriss model,50 nutnber generator was then used to select the layer runs for the Geomatrix model and 100 runs for the void ratio, from which the initial shear modulus of no degradation model). Random number the clay soil layer calculated. Iterative convolution generaton were used throughout to select soil calculations were then performed suitably stiffness and damping properties used in these accounting for depadation in this stiffness with calculations and prende a range of output motion I
cyclic shear stninmg. Sepacate evaluations of die suitably accounting for variability in these response for a carticular to:k input modon were propertes. Each surface specca cornpuced for a made using the various degradation models pven spectralinput was then stored in a data base discussed above, associated with that input. After the surface spectral data was amassed, median and median plus/minus SURFACE IIAZARD CALCULATIONS one sigma spectral accelerations were computed at each frequency of intnest. An example of the For the soil column denned for the site, the results it town in Fig. 6, in which median surface following ?rocedure was then used to obtain spectra are shown for the case of a median rock estimates of the site specific seismic hazards, and outcrop input motion at the 1,000 year return this procedure is shown schematically in Fig. 4. pe:iod. Although spectral values were computed at First, the haard data defined as the bedrock many frequencies, only the results at the six outcrop hazard was obtained for retum periods of frequencies defined for the input spectra are 100 to 1,000,000 years. This data was available in shown.
the form of PGA and spectral accelentions (at frequencies of 1, 2.5, 5,10 and 25 hertz) at Sevent imponant facts can be deduced from orobability fractiles of 15%,50% and 85E The these results. Fint, the comparison between the
?GA hazard data is shown in Fig. 5 for both the LLNL spectra at bedrock and at the ground surface bedrock input as well as at the ground surface show relatively small differences at all frequency using the generic amplification factors from the intervals, even at the lower frequencies associated LLNL study. As may be noted from these data, no with the deep soll column. Secondly, the impact of significant change in peak a:celeration is postulated the soil degndation models assumed for the soils in this approach fro:n bedrock to the ground of the column completely dominate the magnitude surface, even for a soil column as long as 1550 and shape of the computed surface response feet. spectra. For larger degradation values of the Seed-Idriss type, the magnitude cf response at the For each retum period and fnetile, a ground surface is significantly reduced as recommended bedrock spectra was then available compared to the input, particularly at frequencies as the estimate of the seismic input to the soll above 10 hz, and increased at the lower frequencies colurnn. This spectra was considered to be the (below 2.5 hz). The results associated with the definition of the motion of the rock outcrop Geomatrix model, on the other hand, show associated with the site. For any one column signincantly higher surface response at all evaluation, a time history was generated to match frequencies above about I hz. This is primarily this defined bedrock spectra, utilizing a random due to the low damping, particularly at the deeper distribution of tirne phasing of the frequency depths of the soil column, defined in this model.
co:nponents making up the seismic pulse and For higher input accelention levels associated with tnatchbg the specified peak ground accelention by longer return periods, it was found that even this i
" clipping". The frequency range considered in any model indicated shifts in column frequency to one ame history development was up to 50 hertz at lower values due to soil depadadon effects, nus a frequency merement of 0.05 hertz. Pulse changes in spectral shape with retum period can be durations used in the calculations were 20 seconds. significantin these calculations.
SITE SPECIFIC RESULTS ne Geomatrix degradation model was then selected for the surface hazard calculation since it For each of five return periods and three leads to much higher predictions of surface motion inctiles considered in the calculations, a number of than the other Inodels considered, and therefore can hird DOE Natural phenomena Hazards Mitigation Conference - 1991 207 I.
be considered conservative for this site. The Laboratory, October 1990 median output spectral results of these calculations for the 15th, 50th and 85th percentiles spectral (2) C. J. Costantino, E. Heymsfield, Y. T. Gu, definitions of the input motions are plotted m Fig. "' Site Specific Estimates of Surface Ground
- 7. The stectral accelerations shawr in this figure Motions for the HFBR Site at Brookhaven are the average of the 5 and 10 hz iesponses which Topical Report No.
were deemed most important for the structural rish National CE ERC-101, Laboratory",
Earthqua ke Research Center, assessments to be made for the HFBR. As may be Civil Engineering Department, City College noted_ at the higher accelerat.on t levels associated of New York for Brookhaven National with the bedrock inputs, the magnitudes of the Laboratory, /
- ebruary,1991 surface response fall significantly below the initial slope, indicating the nonlinear behavior of the soil [3] J. Xu, A. J. Philippacopoulos, C A. Miller, column at the higher acceleration levels. This result C. J. Costantino, , CARES", NUREG/CR-is even more striking at the higher frequency 5588, vols. I thru 3 Brookhaven National levels. This behavior ca'i be thought of as a Laboratory for U. S. Nuclear Regulatory
" saturation" of the soil column indicating that the Commission, July,1990 column is ne longer abic to transmit the larger spectral accelerationt associated with the higher [4] H. B. Seed and I. M. Idriss, " Soil Moduli input motions. and Damping Factors for Dynamic Response Analyses ', Report No. EERC-70-10,
'the impact of this column saturation on the University of Cahfomia, Berkeley, December surface seismic hazard curves is shown in Fig. 8. 1970 The hazard data plotted is return period as a function of the average spectral acceleration (5 to [5] K. Coppersmith, " Ground Motion Following 10 hz), although rimilar conclusions would be Selecnon of SRS Design Basis Earthquake reached for any other spectral acceleration of and Associated Deterministic Approach",
interest. The generic rock outcrop curves of Fig. 5 Geomatrix Consultants, Draft Final Report, were then converted to the site specific stuface Project No. 1724, for Westinghouse hazard curves of Fig. 8 using the deterministic Savannah River Cornpany. January,1991.
relationship of Fig. 7. The additional variability associated vdth this conversion, including the [6] I. M. Idriss, ." Response of Soft Soil Sites scatter shown, was incorporated into this Dtuing Earthquakes", Proceedings of the H.
calculation. The detaih of this computation are B. Seed Memorial Symposium, Berkeley, presented in (9). The results of Fig. 8 suggest that Califomia,May,1990 at the '.ower input acceleration levels (or shorter retum periods), the site specific hazard is greater (7) H. J. Gibbs and W. G. Holtz, "Research on than would be predicted using the genene soil Determining the Density of Sands by Spoon amplification factors 7redicted with the generic Penetration Testing", Proceedings of the 4th approach. On the other hand, at higher input ICSMFE, vol.1,1957 acceleration levels (longer return periods), the .
surface hazard 9 significantly reduced fmm that [8] " Soil Behavior Under Earthquake Loading predicted from the use of the generic darn Conditions; State of the Art Evaluation of Soil Characteristics For Seismic Response ACKNOWLEDGEMENT Analyses", Agbabian-Jacobsen Associates and Shannon & Wilson Inc. for the U. S.
This work was performed under the auspices of the Atomic Energy Commission, January 1972.
U.S. Dept. of Energy by Brookhaven National Laboratory. [9] YJ. Park, et al, " Seismic Hazard & Fragility of Structures & Cornponents for Use in the REFERENCES PRA for the HFBR", Structural Analysis
, Division, BNL, August 1991 (Drzft)
[1] J. B. Savy, " Seismic Hazard Characterization of the BNL-HFBR Site", UCRL-ID-105148, Lawrence Livermore National Third DOE Natural Phenomena Hazards Mitigation Conference - 1991 208 a